During the Data and AI Night, Eramet was awarded first prize in the “Data & AI Team Projects” category for its predictive maintenance project for the Transgabonais, Gabon’s national railway.
This ceremony, organized annually by Republik Data & IA, rewards innovative projects involving the use of artificial intelligence. A jury composed of experts from major French companies selects the best initiatives in several categories.
It is worth noting that the project was also selected from among 600 entries for the World Summit on Artificial Intelligence organized by the Elysée on February 10 and 11.
AI and predictive maintenance: preventing rather than curing
In Gabon, our subsidiary Setrag operates the country’s only railway, the Transgabonais, under the mandate of the authorities. This railway, nearly 650 kilometers long, connects the western and eastern ends of the country, serving 24 stations. The train transports manganese ore from our subsidiary Comilog, as well as other raw materials and passengers.
“This line, commissioned in the 1980s, was designed to support ten times less traffic than it does today,” explains Jean-Loup Loyer, Chief Data Management & Analytics Officer at Eramet. “Moreover, Gabon’s tropical climate makes the soil wet and unstable, which also causes the track to shift.”
To address this issue, Eramet has developed a predictive maintenance project based on AI. Specifically, the Group has invested in a maintenance-dedicated locomotive. Equipped with sensors, it collects measurement data (rail gap, track shape). For several years, this locomotive has been collecting reliable data every month, allowing us to predict, kilometer by kilometer, future degradation points on the line.
“With this knowledge, we are able to make recommendations to the Setrag maintenance teams,” concludes Jean-Loup Loyer. “This not only improves our operational efficiency but also safety.” A project that impressed the jury of the Data and AI Night for its originality and practical application.
AI and Data at the service of operational performance
The mining sector faces numerous challenges, whether in terms of performance, social responsibility, or safety. This is why Eramet deploys several data-driven projects across all its sites worldwide.
In Gabon, for example, a project aims to automatically evaluate the arrival times of Setrag trains via surveillance cameras on the platform. These cameras can detect the train’s arrival, the number of wagons, and the loading level of the ore wagons. This data helps better plan operations and develop operational improvement action plans.
Also in Gabon, at the Comilog mine, predictive maintenance for mining equipment is being considered, similar to what other mining companies are already doing. Sensors on the vehicles collect operational information, which is then analyzed. For example, it is possible to check that the equipment is not over-consuming fuel, that maintenance operations are properly carried out, and that speed limits are respected. Any discrepancies can lead to awareness or training sessions.
Finally, in Norway, process optimization is also underway at the pyrometallurgical furnaces. At the Eramet Norway site in Kvinesdal, artificial intelligence algorithms allow Eramet’s data teams to provide metallurgists with recommendations to improve the process or reduce the energy consumption of the furnaces.
